This repo contains code to mirror other repos. It also contains the code that is getting mirrored.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

158 lines
5.0 KiB

  1. """This module implements a LALR(1) Parser
  2. """
  3. # Author: Erez Shinan (2017)
  4. # Email : erezshin@gmail.com
  5. from copy import deepcopy, copy
  6. from ..exceptions import UnexpectedCharacters, UnexpectedInput, UnexpectedToken
  7. from ..lexer import Token
  8. from .lalr_analysis import LALR_Analyzer, Shift, Reduce, IntParseTable
  9. from .lalr_puppet import ParserPuppet
  10. ###{standalone
  11. class LALR_Parser(object):
  12. def __init__(self, parser_conf, debug=False):
  13. analysis = LALR_Analyzer(parser_conf, debug=debug)
  14. analysis.compute_lalr()
  15. callbacks = parser_conf.callbacks
  16. self._parse_table = analysis.parse_table
  17. self.parser_conf = parser_conf
  18. self.parser = _Parser(analysis.parse_table, callbacks, debug)
  19. @classmethod
  20. def deserialize(cls, data, memo, callbacks, debug=False):
  21. inst = cls.__new__(cls)
  22. inst._parse_table = IntParseTable.deserialize(data, memo)
  23. inst.parser = _Parser(inst._parse_table, callbacks, debug)
  24. return inst
  25. def serialize(self, memo):
  26. return self._parse_table.serialize(memo)
  27. def parse(self, *args):
  28. return self.parser.parse(*args)
  29. class ParseConf:
  30. __slots__ = 'parse_table', 'callbacks', 'start', 'start_state', 'end_state', 'states'
  31. def __init__(self, parse_table, callbacks, start):
  32. self.parse_table = parse_table
  33. self.start_state = self.parse_table.start_states[start]
  34. self.end_state = self.parse_table.end_states[start]
  35. self.states = self.parse_table.states
  36. self.callbacks = callbacks
  37. self.start = start
  38. class ParserState:
  39. __slots__ = 'parse_conf', 'lexer', 'state_stack', 'value_stack'
  40. def __init__(self, parse_conf, lexer, state_stack=None, value_stack=None):
  41. self.parse_conf = parse_conf
  42. self.lexer = lexer
  43. self.state_stack = state_stack or [self.parse_conf.start_state]
  44. self.value_stack = value_stack or []
  45. @property
  46. def position(self):
  47. return self.state_stack[-1]
  48. def __copy__(self):
  49. return type(self)(
  50. self.parse_conf,
  51. self.lexer, # XXX copy
  52. copy(self.state_stack),
  53. deepcopy(self.value_stack),
  54. )
  55. def copy(self):
  56. return copy(self)
  57. def feed_token(self, token, is_end=False):
  58. state_stack = self.state_stack
  59. value_stack = self.value_stack
  60. states = self.parse_conf.states
  61. end_state = self.parse_conf.end_state
  62. callbacks = self.parse_conf.callbacks
  63. while True:
  64. state = state_stack[-1]
  65. try:
  66. action, arg = states[state][token.type]
  67. except KeyError:
  68. expected = {s for s in states[state].keys() if s.isupper()}
  69. raise UnexpectedToken(token, expected, state=state, puppet=None)
  70. assert arg != end_state
  71. if action is Shift:
  72. # shift once and return
  73. assert not is_end
  74. state_stack.append(arg)
  75. value_stack.append(token)
  76. return
  77. else:
  78. # reduce+shift as many times as necessary
  79. rule = arg
  80. size = len(rule.expansion)
  81. if size:
  82. s = value_stack[-size:]
  83. del state_stack[-size:]
  84. del value_stack[-size:]
  85. else:
  86. s = []
  87. value = callbacks[rule](s)
  88. _action, new_state = states[state_stack[-1]][rule.origin.name]
  89. assert _action is Shift
  90. state_stack.append(new_state)
  91. value_stack.append(value)
  92. if is_end and state_stack[-1] == end_state:
  93. return value_stack[-1]
  94. class _Parser:
  95. def __init__(self, parse_table, callbacks, debug=False):
  96. self.parse_table = parse_table
  97. self.callbacks = callbacks
  98. self.debug = debug
  99. def parse(self, lexer, start, value_stack=None, state_stack=None):
  100. parse_conf = ParseConf(self.parse_table, self.callbacks, start)
  101. parser_state = ParserState(parse_conf, lexer, state_stack, value_stack)
  102. return self.parse_from_state(parser_state)
  103. def parse_from_state(self, state):
  104. # Main LALR-parser loop
  105. try:
  106. token = None
  107. for token in state.lexer.lex(state):
  108. state.feed_token(token)
  109. token = Token.new_borrow_pos('$END', '', token) if token else Token('$END', '', 0, 1, 1)
  110. return state.feed_token(token, True)
  111. except UnexpectedInput as e:
  112. try:
  113. e.puppet = ParserPuppet(self, state, state.lexer)
  114. except NameError:
  115. pass
  116. raise e
  117. except Exception as e:
  118. if self.debug:
  119. print("")
  120. print("STATE STACK DUMP")
  121. print("----------------")
  122. for i, s in enumerate(state.state_stack):
  123. print('%d)' % i , s)
  124. print("")
  125. raise
  126. ###}